I'd like to introduce you to https://Panda.chat — an end-to-end encrypted LLM designed to protect your privacy. It runs Deepseek R1 & Llama 4 Scout.

These days, we're asking ChatGPT for medical, legal, and mental advice. But some AI services aren't designed to protect our privacy. Rather, they're designed to log your inputs & train the model for better performance.

We built https://Panda.chat to solve this.

Panda runs AI models in Trusted Execution Environment (TEE) inside of the GPU, enabling end-to-end encryption from your device to TEE. So no one—including us—can see your input/response. We can also prove your conversation was fully protected with a code hash posted on Ethereum.

Try http://panda.chat today! Enjoy the fully end-to-end encrypted LLM experiences. Would love to get your feedback on this :)

Source code (yup, it's open-sourced!):

- https://github.com/testinprod-io/panda-web

- https://github.com/testinprod-io/panda-vllm-proxy

Docs:

- One Pager: https://testinprod.notion.site/Panda-One-Pager-1ff8fc57f5468...

- Technical FAQ: https://testinprod.notion.site/Panda-Technical-FAQ-2018fc57f...